Chronilogix CAIHL draft report

Chronilogix: draft HugoScore review

Review date: September 28, 2026. Status: unscored AI-assisted draft; publication authorized by Hugo Campos.

1. Executive Summary

Chronilogix offers patient-facing AI coaching with potential to support reflection and personally meaningful goals. Its institutional sales model emphasizes adherence and claims reduction. The resulting agency posture is Mixed, with patient authority in specific deployments unverified. Public disclosures contain material conflicts about model training and the service's clinical role.

2. CAIHL Question

Who does this AI serve? Patients use the coaching; employers, plans, brokers and product vendors are prominent buyers. These interests can coincide, but patient control cannot be inferred from an empathetic conversational style. Broker offering

3. What The Service Does

• Tool/vendor: Chronilogix / Chronilogix, Inc.
• URL: https://chronilogix.com/
• Category: Condition-specific coaching and care AI; mental health coaching.
• Coaches: Roni for diabetes/chronic care; Millie for mental health/mood.
• Primary users: patients or members; organizational purchasers shape deployments.
• Hosting/control: vendor-hosted, with partner embedding and a consumer app listing. Direct consumer and institutional workflows require separate assessment.

Sources: Product, App Store.

4. Patient-Impact Pathway

AI conversations may influence self-management, care-seeking and health decisions. The tool passes the AI, health and patient-impact scope gates. Potential benefits include accessible reflection and goal setting; potential constraints include institution-selected goals and unclear control over sharing and AI-created interpretations. These are plausible pathways, not observed app behavior.

5. Evidence Table

Training disclosures conflict

The FAQ excludes conversation use for model improvement. Privacy policy III.B permits training with primarily aggregated/de-identified content; V preserves anonymized improvement data after deletion. This establishes inconsistent disclosures, not proof of actual training. FAQ, Privacy

Clinical role needs clarification

The FAQ denies clinician replacement; the broker page proposes replacing up to 80% of human coaching sessions. Terms 2.1–2.3 define self-help, exclude treatment and emergency support, and describe hotline redirection. Coaching replacement and clinician replacement are not identical, but the practical boundary needs explanation. FAQ, Brokers, Terms

Headline outcome is evidence for another intervention

The homepage attributes its 58% example to the Diabetes Prevention Program. NIH confirms the intensive lifestyle intervention result. This supports the general rationale, not the effectiveness of Chronilogix; the site does not label it a Chronilogix trial. Homepage, NIDDK

Scientific association has substance but limited evidentiary reach

Minnesota independently lists Kenneth Resnicow as faculty. That confirms academic standing, not product performance. University directory

6. Mixed HugoScore Profile

Agency posture

Mixed

Classification

Hybrid; prominent institution-directed commercial deployment

Axis position

Unscored; no cross-deployment numerical placement justified

Patient visibility

Yes in reviewed public materials; embedded deployment disclosure untested

Choice and refusal

Partial: app lists adjustable goals; refusal without benefit consequences unknown

Correction and challenge

Partial: policy provides data correction rights; correction of AI interpretations untested

Action support

Potentially: reflection and goal setting; advocacy against institutional decisions not established

Equity burden

Cultural tailoring and broader access are claimed; accessibility and subgroup performance unverified

Data governance

Conversation storage; conditional access/portability rights; deletion provisions with derivative-data exception; conflicting training statements

Clinical boundaries

Partial/inconsistent across promotional and contractual material

Evaluation ownership

Vendor/scientific leadership visible; patient partnership in defining evaluation outcomes not established

Sources: Product, App Store, Privacy, About.

7. Key Unknowns

Which training statement governs each deployment? Can patients decline sponsor goals, inspect and correct inferred profiles, control clinician summaries, and leave without consequences? Who receives identifiable conversations? What are the model/provider arrangements and human escalation response times? Are product-specific trials, adverse-event results and independent MI-fidelity evaluations available? Targeted searches did not locate a peer-reviewed evaluation of the current Roni/Millie products; that is not proof none exists.

8. Patient Agency Interpretation

The decisive behavioral test is whether the coach helps a patient pursue a goal that conflicts with the purchaser's preferred outcome. Empathy and engagement alone cannot establish that authority. Cost reduction can benefit patients; it does not itself demonstrate patient alignment. Risks identified here concern opacity, expectation mismatch and possible institutional steering, not established harm.

9. Publication Recommendation

Publication authorized as an unscored, source-backed AI-assisted draft with deployment qualifications. Training and clinical-scope questions remain unresolved. Publication is not endorsement or comprehensive human verification.

Confidence: medium for documentary findings; low for actual behavior and efficacy. Review provenance: OpenAI Codex / GPT-6 (exact model variant not recorded); CAIHL and current HugoScore framework. Hugo Campos authorized publication; comprehensive human verification has not occurred. Public-source review only; no account, health-data upload, hands-on testing, security audit or vendor interview. Evidence access date: September 28, 2026.